asymmetrical waveform?

[Paul[_19_]]

Hi,

I have received some files (spoken word) from a recording studio some
of which seem to be not quite symmetrical. Is it a DC offset or
something? Some of the spikes are +ve going (assuming up is +ve) and so
normalizing really doesn't work. Some other files recorded at the same
studio on a different day are symmetrical.

What would cause such a thing? A faulty mic or amp or something?

Thanks,

Paul

[Don Pearce[_3_]]

On Thu, 1 Oct 2009 20:57:15 +1000, Paul >
wrote:

>Hi,
>
>I have received some files (spoken word) from a recording studio some
>of which seem to be not quite symmetrical. Is it a DC offset or
>something? Some of the spikes are +ve going (assuming up is +ve) and so
>normalizing really doesn't work. Some other files recorded at the same
>studio on a different day are symmetrical.
>
>What would cause such a thing? A faulty mic or amp or something?
>
>Thanks,
>
>Paul
>

No faults, that is simply how voices are sometimes. There is probably
no DC offset. What you will find is that the taller bits are narrower
than the shorter bits, leaving the area (which is what determines
symmetry) equal on both sides of the line.

d

[Ty Ford]

On Thu, 1 Oct 2009 06:57:15 -0400, Paul wrote
(in article >):

> Hi,
>
> I have received some files (spoken word) from a recording studio some
> of which seem to be not quite symmetrical. Is it a DC offset or
> something? Some of the spikes are +ve going (assuming up is +ve) and so
> normalizing really doesn't work. Some other files recorded at the same
> studio on a different day are symmetrical.
>
> What would cause such a thing? A faulty mic or amp or something?
>
> Thanks,
>
> Paul
>
>

Not necessarily. Voices are usually asymmetrical. I have more negative peaks
than positive. In fact, that seems to add to my voice cutting through a bit
more when I do VO. Of course, if your rig has a polarity flip somewhere, then
they go positive,

Regards,

Ty Ford


--Audio Equipment Reviews Audio Production Services
Acting and Voiceover Demos http://www.tyford.com
Guitar player?:http://www.youtube.com/watch?v=yWaPRHMGhGA

[Don Pearce[_3_]]

On Thu, 1 Oct 2009 08:20:53 -0400, Ty Ford >
wrote:

>On Thu, 1 Oct 2009 06:57:15 -0400, Paul wrote
>(in article >):
>
>> Hi,
>>
>> I have received some files (spoken word) from a recording studio some
>> of which seem to be not quite symmetrical. Is it a DC offset or
>> something? Some of the spikes are +ve going (assuming up is +ve) and so
>> normalizing really doesn't work. Some other files recorded at the same
>> studio on a different day are symmetrical.
>>
>> What would cause such a thing? A faulty mic or amp or something?
>>
>> Thanks,
>>
>> Paul
>>
>>
>
>Not necessarily. Voices are usually asymmetrical. I have more negative peaks
>than positive. In fact, that seems to add to my voice cutting through a bit
>more when I do VO. Of course, if your rig has a polarity flip somewhere, then
>they go positive,
>
>Regards,
>
>Ty Ford

I'm surprised, because one effect of peakiness in one direction is
that you can't bring the volume up as far before you start clipping
one end. Symmetrical voices are best for loudness.

d

[Scott Dorsey]

Paul > wrote:
>
>I have received some files (spoken word) from a recording studio some
>of which seem to be not quite symmetrical. Is it a DC offset or
>something? Some of the spikes are +ve going (assuming up is +ve) and so
>normalizing really doesn't work. Some other files recorded at the same
>studio on a different day are symmetrical.

Voices are like that.

>What would cause such a thing? A faulty mic or amp or something?

No, that's normal. I'd wonder what they did to the other tracks to make
them symmetric. Could be compression. Could be a different person speaking.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Don Pearce > wrote:
>
>I'm surprised, because one effect of peakiness in one direction is
>that you can't bring the volume up as far before you start clipping
>one end. Symmetrical voices are best for loudness.

That's why we run them through the phase rotator and then compress the
crap out of them so they are nice and symmetric afterward. Not very natural,
but that's what the announcers want...
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Don Pearce[_3_]]

On 1 Oct 2009 10:35:23 -0400, (Scott Dorsey) wrote:

>Don Pearce > wrote:
>>
>>I'm surprised, because one effect of peakiness in one direction is
>>that you can't bring the volume up as far before you start clipping
>>one end. Symmetrical voices are best for loudness.
>
>That's why we run them through the phase rotator and then compress the
>crap out of them so they are nice and symmetric afterward. Not very natural,
>but that's what the announcers want...
>--scott

You know you just said that out loud...?

d

[Mike Rivers]

Paul wrote:

> I have received some files (spoken word) from a recording studio some of
> which seem to be not quite symmetrical. Is it a DC offset or something?

No, the human voice, and many musical instruments, are like that. Now that
you have the capability to view waveforms, you can see all sorts of
things that
seem "wrong" about audio.

Do they sound OK? That's what counts.

[Arkansan Raider]

Don Pearce wrote:
> On 1 Oct 2009 10:35:23 -0400, (Scott Dorsey) wrote:
>
>> Don Pearce > wrote:
>>> I'm surprised, because one effect of peakiness in one direction is
>>> that you can't bring the volume up as far before you start clipping
>>> one end. Symmetrical voices are best for loudness.
>> That's why we run them through the phase rotator and then compress the
>> crap out of them so they are nice and symmetric afterward. Not very natural,
>> but that's what the announcers want...
>> --scott
>
> You know you just said that out loud...?
>
> d

Shhhhhh... I won't tell. ;)

---Jeff

[Richard Crowley]

"Paul" wrote ...
> I have received some files (spoken word) from a recording studio some of
> which seem to be not quite symmetrical. Is it a DC offset or something?
> Some of the spikes are +ve going (assuming up is +ve) and so normalizing
> really doesn't work. Some other files recorded at the same studio on a
> different day are symmetrical.
>
> What would cause such a thing? A faulty mic or amp or something?

Completely normal, IME.
Assuming it *sounds* right to the ear.

[Arny Krueger]

"Paul" > wrote in message
u

> I have received some files (spoken word) from a recording
> studio some of which seem to be not quite symmetrical. Is
> it a DC offset or something? Some of the spikes are +ve
> going (assuming up is +ve) and so normalizing really
> doesn't work. Some other files recorded at the same
> studio on a different day are symmetrical.

> What would cause such a thing? A faulty mic or amp or
> something?

No, all it takes is the presence of even harmonics in the input signal for
that signal to be asymmetrical.

[Tobiah]

> No, all it takes is the presence of even harmonics in the input signal for
> that signal to be asymmetrical.

Why would that be? Mathematically, the harmonics do
not present asymmetry.

[William Sommerwerck]

>> No, all it takes is the presence of even harmonics
>> in the input signal for that signal to be asymmetrical.

> Why would that be? Mathematically, the harmonics do
> not present asymmetry.

You need to check up on Fourier analysis. Broadly speaking, asymmetery in a
repeating waveform is caused by even-order harmonics. That's a fact.

If you don't believe this, draw a sine wave and its second harmonic. Then
add them. See?

[Don Pearce]

On Thu, 01 Oct 2009 12:03:18 -0700, Tobiah > wrote:

>> No, all it takes is the presence of even harmonics in the input signal for
>> that signal to be asymmetrical.
>
>Why would that be? Mathematically, the harmonics do
>not present asymmetry.

Think about what is happening. An extra waveform is being added that
has twice the frequency. That means that if it adds a positive peak to
the positive half cycle, it will also add a positive peak to the
following negative half cycle. The result is that the positive half
cycle is stretched upwards, and the negative half cycle is squashed
upwards. The result is asymmetry.

This goes for any even harmonic.

d
--
Pearce Consulting
http://www.pearce.uk.com

[[email protected]]

While it may be true that some sounds (voices, whatever) just
are like this, I believe that it is also possible to produce
these sort of asymmetrical waveforms by some recording methods
(but I don't know exactly how). The reason is, about 10 years
ago I recorded a large brass line in a number of settings.
One of the settings was in a large circle, and I recorded
from the center. That particular recording, with the same
brass line, the same tune, same horns, same players, turned out
extremely asymmetrical, whereas the other recordings of the
same group were symmetrical. The only change was the venue
and mic placement.

Unfortunately, that recording was the one with the best
performance and the one they wanted on the CD. It was a
bear to master and get much volume out of.

[Scott Dorsey]

> wrote:
>While it may be true that some sounds (voices, whatever) just
>are like this, I believe that it is also possible to produce
>these sort of asymmetrical waveforms by some recording methods
>(but I don't know exactly how). The reason is, about 10 years
>ago I recorded a large brass line in a number of settings.
>One of the settings was in a large circle, and I recorded
>from the center. That particular recording, with the same
>brass line, the same tune, same horns, same players, turned out
>extremely asymmetrical, whereas the other recordings of the
>same group were symmetrical. The only change was the venue
>and mic placement.

There is _nothing_ more asymmetric than a brass instrument. They have
gargantuan peaks on the positive-going side, and very narrow ones.

I would wonder what you did with the other recordings to make them
symmetric. Most common thing is tape limiting.... if you are using
VU metering, you need to leave outrageous amounts of headroom for
brass instruments or you will severely limit the peaks.

It's possible if the bells were pointing way away from the microphones
that you might be losing a lot of that. And of course once the ambient
sound bounces around a couple times it gets smoothed out a lot.

>Unfortunately, that recording was the one with the best
>performance and the one they wanted on the CD. It was a
>bear to master and get much volume out of.

The good news is that you can also limit the crap out of brass instruments
and people don't notice so much. They get less blatty and more mellow, but
mostly people like that.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[William Sommerwerck]

> There is _nothing_ more asymmetric than a brass instrument.
> They have gargantuan peaks on the positive-going side, and
> very narrow ones.

Piano, perhaps, but you otherwise took the words right out of my mouth.
Brass instruments are downright nasty when it comes to avoiding clipping in
digital recordings, and saturation in analog.

[Ty Ford]

On Thu, 1 Oct 2009 08:24:55 -0400, Don Pearce wrote
(in article >):

> On Thu, 1 Oct 2009 08:20:53 -0400, Ty Ford >
> wrote:
>
>> On Thu, 1 Oct 2009 06:57:15 -0400, Paul wrote
>> (in article >):
>>
>>> Hi,
>>>
>>> I have received some files (spoken word) from a recording studio some
>>> of which seem to be not quite symmetrical. Is it a DC offset or
>>> something? Some of the spikes are +ve going (assuming up is +ve) and so
>>> normalizing really doesn't work. Some other files recorded at the same
>>> studio on a different day are symmetrical.
>>>
>>> What would cause such a thing? A faulty mic or amp or something?
>>>
>>> Thanks,
>>>
>>> Paul
>>>
>>>
>>
>> Not necessarily. Voices are usually asymmetrical. I have more negative
>> peaks
>> than positive. In fact, that seems to add to my voice cutting through a bit
>> more when I do VO. Of course, if your rig has a polarity flip somewhere,
>> then
>> they go positive,
>>
>> Regards,
>>
>> Ty Ford
>
> I'm surprised, because one effect of peakiness in one direction is
> that you can't bring the volume up as far before you start clipping
> one end. Symmetrical voices are best for loudness.
>
> d

That's why the phase rotator was invented. However, that's a distortion of
the original waveform most purists can not abide.

AM radio loved it because coverage area is partly determined by the amount of
modulation.

Regards,

Ty Ford


--Audio Equipment Reviews Audio Production Services
Acting and Voiceover Demos http://www.tyford.com
Guitar player?:http://www.youtube.com/watch?v=yWaPRHMGhGA

[Ty Ford]

On Thu, 1 Oct 2009 19:02:00 -0400, Scott Dorsey wrote
(in article >):

> > wrote:
>> While it may be true that some sounds (voices, whatever) just
>> are like this, I believe that it is also possible to produce
>> these sort of asymmetrical waveforms by some recording methods
>> (but I don't know exactly how). The reason is, about 10 years
>> ago I recorded a large brass line in a number of settings.
>> One of the settings was in a large circle, and I recorded
>> from the center. That particular recording, with the same
>> brass line, the same tune, same horns, same players, turned out
>> extremely asymmetrical, whereas the other recordings of the
>> same group were symmetrical. The only change was the venue
>> and mic placement.
>
> There is _nothing_ more asymmetric than a brass instrument. They have
> gargantuan peaks on the positive-going side, and very narrow ones.
>
> I would wonder what you did with the other recordings to make them symmetric.

> Most common thing is tape limiting.... if you are using VU metering, you
> need to leave outrageous amounts of headroom for brass instruments or you
> will severely limit the peaks.

'persactly!

Regards,

Ty Ford


--Audio Equipment Reviews Audio Production Services
Acting and Voiceover Demos http://www.tyford.com
Guitar player?:http://www.youtube.com/watch?v=yWaPRHMGhGA

[Peter Larsen[_3_]]

Paul wrote:

> Hi,

> I have received some files (spoken word) from a recording studio some
> of which seem to be not quite symmetrical. Is it a DC offset or
> something? Some of the spikes are +ve going (assuming up is +ve) and
> so normalizing really doesn't work. Some other files recorded at the
> same studio on a different day are symmetrical.

> What would cause such a thing?

The actual generation of the sound, ie. mechanical properties of the larynx,
and the natural unlinearity of the air. 2'order distortion later in the
recording process may also matter. There is a slight difference of tonal
perception of sounds that are asymmetric ""downwards"" vs. ""upwards"" when
you see the curveform on your screen, kinda like ""flat"" vs. ""sharp"",
also a bit of ""closed"" vs. ""open"". There keeps being stuff around that
breaks absolute polarity, imo. it is preferable to verify actual equipment
used, also because it can have implications for how acoustic cross-talk -
bleed - works in a mix.

The stuff that shouldn't be there should definitely have the same polarity
in all the channels/tracks it shouldn't be on lest it creates haywire with
your ambience.

> A faulty mic or amp or something?

Sound just is like that, but microphones have their say on this too, my
general finding is that recordings made with omni's are less asymmetric than
those made with directional microphones. Dynamic microphones come with
additional sonic challenges due to the acoustic equalizers involved in
shaping their frequency response.

Asymmetry may differ in case the phase response is influenced via minimum
phase filtering, such as a conventional high pass filter. Non-minimum phase,
aka "phase linear" filtering is on the market, generally most aberrrations
to eq away are minimum phase by nature, so may be less of an advantage than
sales talks sugggest.

> Paul

Kind regards

Peter Larsen